The Cancer Genome Atlas (TCGA) via the Genomic Data Commons
National Cancer Institute and National Human Genome Research Institute
Tumors from about 11,000 patients, profiled every molecular way available in the 2010s.
TCGA molecularly characterized tumors and matched normal tissue for 33 cancer types: mutations, copy number, gene expression, methylation, protein levels, and pathology slides, plus clinical data. The NCI Genomic Data Commons now hosts it, harmonized to a current reference genome. Open-tier files cover expression, clinical data, and somatic mutation calls; germline variants and raw reads require dbGaP approval.
Access
- Cost
- FreeOpen tier downloadable without an account; controlled tier needs a dbGaP data access request.
- License
- Provider's own termsNIH GDC data access policies; controlled data under dbGaP terms
- Sign-up
- No sign-up
- How you get it
- API and Bulk downloadMAF, TSV, BAM, SVS images
What's in it
- One row is
- One case (patient), with multiple data files per sample and data type
- Covers
- Mostly U.S. tissue source sites
- Years
- 2006-2018 collection; harmonized releases continue
- Updated
- Irregular
- Size
- About 11,000 cases across 33 cancer types; over 2 petabytes including raw data
Quality
The landmark multi-omic cancer data set, harmonized and widely re-analyzed. Clinical follow-up is uneven and the cohort is not population-based.
- Clinical outcome data are incomplete
- Primary untreated tumors only, few metastases
- Raw data are controlled access
Connects to
- GTEx Portal: Normal-tissue expression baseline for tumor comparisons
- dbGaP: Gatekeeper for germline and raw sequencing data
- Human Protein Atlas: Its pathology section links protein staining to TCGA survival data
Questions it could answer
- Which genes are mutated across many cancer types and which are tissue-specific?
- Does expression of a gene predict survival within a cancer type?